Method, system and program product for predicting computer system resource consumption

• US 20060253855A1
• Filed: 05/04/2005
• Published: 11/09/2006
• Est. Priority Date: 05/04/2005
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Grabarnik, Genady, Kozalov, Lev, Shwartz, Larisa, Kim, Moon Ju
• Granted Patent: US 7,831,976 B2
• Time in Patent Office: Days
• US Class Current: 718/104
• CPC Class Codes: G06F 11/3404 for parallel or distributed programming
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G06F 11/3414 Workload generation, e.g. scripts, playback
• CPC Class Codes: G06F 11/3447 Performance evaluation by modeling
• CPC Class Codes: G06F 11/3466 Performance evaluation by tracing or monitoring
• CPC Class Codes: G06F 2209/5019 Workload prediction
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: G06F 9/5072 Grid computing

Abstract

Under the present invention, a computer work gradient matrix is built for each computer system that is interconnected in an environment. For each computer system for which resource consumption is desired to be predicted (e.g., “target” computer systems), a transition work cocycle is generated and provided to a master computer system. A set of task work paths will be constructed for the master computer system. Thereafter, an initial resource consumption value can be computed on the master computer system based on the computer work gradient matrix and the set of task work paths for the master computer system as well as a current background loading level that is being experienced by the target computer system. Then, using the initial resource consumption and the transition work cocycle for the target computer system(s), a resource consumption of the target computer system(s) can be predicted on the master computer system.

Claims

1. A method for predicting a resource consumption for one computer system on another computer system, comprising:

obtaining a computer work gradient matrix for each of a first computer system and a second computer system;

obtaining, on the second computer system, a transition work cocycle for the first computer system, wherein the transition work cocycle is based on the computer work gradient matrices and maps a work load tensor of the first computer system to a work load tensor of the second computer system;

constructing a set of task work paths for the second computer system;

computing an initial resource consumption on the second computer system based on the computer work gradient matrix of the second computer system, the set of task work paths, and a current background loading level of the first computer system; and

predicting, on the second computer system, a resource consumption of the first computer system based on the initial resource consumption computed on the second computer system and the transition work cocycle for the first computer system.

2. The method of claim 1, wherein the computer work gradient matrices set forth consumption levels of a set of resources of the first computer system and the second computer system in response to a set of background loading levels and a set of elemental work increments applied to the first computer system and the second computer system.

3. The method of claim 2, further comprising:

applying the set of background loading levels to the first computer system and the second computer system;

applying the set of elemental work increments to the first computer system and the second computer system;

monitoring the consumption levels of the set of resources for the first computer system and the second computer system in response to the applied set of background loading levels and set of elemental work increments; and

building the computer work gradient matrix for each of the first computer system and the second computer system based on the monitoring.

4. The method of claim 2, wherein each of the set of task work paths characterize consumption levels for one of the set of resources in response to a computing task applied to the second computer system at a distinct background loading level.

5. The method of claim 4, wherein the constructing step comprises:

applying the distinct background loading level to the second computer system;

applying the computing task to the second computer system;

monitoring the consumption levels for the set of resources in response to the distinct background loading level and the computing task; and

constructing a set of task work paths for the set of resources based on the monitoring.

6. The method of claim 2, wherein the set of resources comprise memory utilization and CPU utilization.

7. The method of claim 1, wherein the computing step comprises:

obtaining the current background loading level of the first computer system;

accessing the computer work gradient matrix and the set of task work paths for the second computer system; and

computing the initial resource consumption using the computer work gradient matrix and the set of task work paths for the second computer system, as well as the current background loading level of the first computer system.

19. The program product of claim 1 S, wherein the constructing step comprises:

applying the distinct background loading level to the second computer system;

applying the computing task to the second computer system;

monitoring the consumption levels for the set of resources in response to the distinct background loading level and the computing task; and

constructing a set of task work paths for the set of resources based on the monitoring.

8. A system for predicting a resource consumption for one computer system on another computer system, comprising:

a system for obtaining a computer work gradient matrix for each of a first computer system and a second computer system;

a system for obtaining, on the second computer system, a transition work cocycle for the first computer system, wherein the transition work cocycle is based on the computer work gradient matrices and maps a work load tensor of the first computer system to a work load tensor of the second computer;

a system for constructing a set of task work paths for the second computer system;

a system for computing an initial resource consumption on the second computer system based on the computer work gradient matrix of the second computer system, the set of task work paths, and a current background loading level of the first computer system; and

a system for predicting, on the second computer system, a resource consumption of the first computer system based on the initial resource consumption computed on the second computer system and the transition work cocycle for the first computer system to the second computer system.

9. The system of claim 8, wherein the computer work gradient matrices set forth consumption levels of a set of resources of the first computer system and the second computer system in response to a set of background loading levels and a set of elemental work increments applied to the first computer system and the second computer system.

10. The system of claim 9, further comprising:

a system for applying the set of background loading levels to the first computer system and the second computer system;

a system for applying the set of elemental work increments to the first computer system and the second computer system;

a system for monitoring the consumption levels of the set of resources for the first computer system and the second computer system in response to the applied set of background loading levels and set of elemental work increments; and

a system for building the computer work gradient matrix for each of the first computer system and the second computer system based on the monitoring.

11. The system of claim 9, wherein each of the set of task work paths characterize consumption levels for one of the set of resources in response to a computing task applied to the second computer system at a distinct background loading level, and wherein the transition work cocycle further maps the set of task work paths for the second computer system to the first computer system.

12. The system of claim 11, further comprising:

a system for applying the distinct background loading level to the second computer system;

a system for applying the computing task to the second computer system;

a system for monitoring the consumption levels for the set of resources in response to the distinct background loading level and the computing task; and

a system for constructing the set of task work paths for the set of resources based on the monitoring.

13. The system of claim 9, wherein the set of resources comprise memory utilization and CPU utilization.

14. The system of claim 8, wherein the system for computing the initial resource consumption comprises:

a system for obtaining the current background loading level of the first computer system;

a system for accessing the computer work gradient matrix and the set of task work paths for the second computer system; and

a system for computing the initial resource consumption by using the computer work gradient matrix and the set of task work paths for the second computer system, as well as the current background loading level of the first computer system.

15. A program product stored on a computer readable medium for predicting a resource consumption for one computer system on another computer system, the computer readable medium comprising program code for performing the following steps:

obtaining a computer work gradient matrix for each of a first computer system and a second computer system;

obtaining, on the second computer system, a transition work cocycle for the first computer system, wherein the transition work cocycle is based on the computer work gradient matrices and maps a work load tensor of the first computer system to a work load tensor of the second computer system;

constructing a set of task work paths for the second computer system;

computing an initial resource consumption on the second computer system based on the computer work gradient matrix of the second computer system, the set of task work paths, and a current background loading level of the first computer system; and

predicting, on the second computer system, a resource consumption of the first computer system based on the initial resource consumption computed on the second computer system and the transition work cocycle for the first computer system.

16. The program product of claim 15, wherein the computer work gradient matrices set forth consumption levels of a set of resources of the first computer system and the second computer system in response to a set of background loading levels and a set of elemental work increments applied to the first computer system and the second computer system.

17. The program product of claim 16, further comprising:

applying the set of background loading levels to the first computer system and the second computer system;

applying the set of elemental work increments to the first computer system and the second computer system;

monitoring the consumption levels of the set of resources for the first computer system and the second computer system in response to the applied set of background loading levels and set of elemental work increments; and

building the computer work gradient matrix for each of the first computer system and the second computer system based on the monitoring.

18. The program product of claim 16, wherein each of the set of task work paths characterize consumption levels for one of the set of resources in response to a computing task applied to the second computer system at a distinct background loading level, and wherein the transition work cocyle further maps the set of task work paths for the second computer system to the first computer system.

20. The program product of claim 16, wherein the set of resources comprise memory utilization and CPU utilization.

21. The program product of claim 15, wherein the computing step comprises:

obtaining the current background loading level of the first computer system;

accessing the computer work gradient matrix and the set of task work paths for the second computer system; and

computing the initial resource consumption using the computer work gradient matrix and the set of task work paths for the second computer system, as well as the current background loading level of the first computer system.

22. A method for deploying an application for predicting a resource consumption for one computer system on another computer system, comprising:

providing a computer infrastructure being operable to;

obtain a computer work gradient matrix for each of a first computer system and a second computer system;

obtain, on the second computer system, a transition work cocycle for the first computer system, wherein the transition work cocycle is based on the computer work gradient matrices and maps a work load tensor of the first computer system to a work load tensor of the second computer system;

construct a set of task work paths for the second computer system;

compute an initial resource consumption on the second computer system based on the computer work gradient matrix of the second computer system, the set of task work paths, and a current background loading level of the first computer system; and

predict, on the second computer system, a resource consumption of the first computer system based on the initial resource consumption computed on the second computer system and the transition work cocycle for the first computer system.